Better roads in Belgium – automated road inspection using artificial intelligence

Contact person: Steve Vanlanduit

Belgium has one of the most dense road networks in the world. For every square kilometre there are on average more than 5 kilometres of road. Unfortunately, the state of the roads in Belgium is quite poor. Holes and cracks in the road surface can potentially lead to damage of your car or even result in accidents (because of abrupt steering manoeuvres). Continuous road inspection is crucial in order to be able to timely repair and maintain the Belgian roads. At this moment, inspection is done visually by road inspectors who walk or drive around in a vehicle while writing down the type of damage they observe.

At the InViLab research group of the university of Antwerp they came up with an alternative inspection procedure. Images of the road surface are taken by using a smartphone that is installed on the dashboard of a car. Then, artificial intelligence is used on each of the recorded images to automatically detect the presence of damage and to identify the type of damage. This way, the condition of the road network in a small village can be inspected in just a day time. The method can be used by local authorities to make an inventory of the damage and to plan maintenance and repairs.
Image with damage classification